.- help for ^prcompw^ (STB-47: sg101) .- pairwise comparisons of means, including the Tukey wsd method ------------------------------------------------------------- ^prcompw^ [varlist] ^aweights^ and ^fweights^ are allowed; see help @weights@. Description ----------- ^prcomp^ performs pairwise comparisons of the means of a variable (Y), grouped by the values of a second variable (X). By default, the output includes a table of Y means and their standard errors, and a table of confidence intervals for the difference in Y mean between all pairs of X values, with individual confidence level determined by the system macro ^S_level^. Optionally, prcomp will switch from confidence intervals to significance tests, graph the table of confidence intervals or tests, and set the simultaneous confidence (significance) level using Tukey's wsd method. By default, prcomp assumes homogeneity of variance in Y and estimates that var- iance from a one-way ANOVA of Y classified by X. An external estimate of error variability can be supplied instead, or comparisons can be made without assum- ing homogeneity of variance, using the Welch-Satterthwaite approach. Controls -------- ^Confidence Intervals^ displays confidence intervals for all pairwise differences in Y means, using the settings selected by the other dialog controls ^Significance Tests^ shows significance tests of no difference in Y mean between each pair of X values using the settings given by the other dialog controls ^Refresh^ forces recalculation of all quantities used for the tests or confidence intervals: Y means, differences in means, confidence interval half-widths, etc. These quantities are saved in Stata's memory and, ordinarily, are re- computed only as necessary (e.g., when the choice of Y, X, confidence level etc. changes). This option exists to insure, should doubt arise, that the correct quantities are being displayed ^Response (Y)^ chooses the response variable (Y) whose means are to be compared ^Groups (X)^ chooses the variable (X) whose values classify the observations ^a-^ or ^f-weight^ chooses between ^aweights^ and ^fweights^; ignored if the edit box below is empty ^^ supplies an ^if^ condition for selecting observations to be included; enter the condition in the box below, without "if" ^^ supplies an ^in^ range for selecting observations to be included; enter the range in the box below, without "in" ^Conf. level^ sets the simultaneous confidence (significance) if the ^Simult.^ box (below) is checked, otherwise the individual level. The default value is the current setting of ^S_level^. Other values can be entered in the box at the left as proportions (.98333) or as percentages (98.333) ^Simult.^ if checked, uses quantiles of the studentized range to control the si- multaneous confidence (significance); the default is to control individual levels with student t quantiles. This option yields the Tukey wsd method; it cannot be combined with the ^Unequal SDs^ option (below), and it requires that the @qsturng@ command be installed ^Unequal SDs^ specifies that standard errors for differences in Y means be esti- mated without a homogeneity of variance assumption using the Welch standard error and Satterthwaite's degrees of freedom (see the ^unequal^ option of the @ttest@ command) ^ANOVA Pooled SD^ uses the RMSE of a one-way ANOVA of Y classified by X to esti- mate the error standard deviation, assumed to be constant over levels of X ^Other^ if checked uses the contents of the edit boxes below to provide an esti- mate of the error standard deviation, assumed to be constant over levels of X. Both the ^Sigma^ and the ^d.f.^ boxes must be filled ^Display order^ controls how values of X are arranged in listings of means and in listings or graphs of confidence intervals (tests). ^Natural^ orders levels of X by their numeric values, ^Labels^ by the value labels currently assign- ed to X, and ^Means^ arranges X so that the means of Y are in order. For each of these choices the order is ascending; the same options with the initial letter in lower case yield descending order ^Use X value labels^ if checked uses the value labels currently assigned to X to identify levels of X in tables and graphs; otherwise, the numerical values of X are used ^ANOVA summary table^ if checked shows the one-way ANOVA used (by default) to ob- an estimate of the error standard deviation ^Table of Y means^ if checked displays a table of group sizes, Y means, and their standard errors ^Table of intervals^ if checked displays a table of confidence intervals for the difference in Y mean between all pairs of X values, or a table of signifi- cance tests for no difference in Y mean between all pairs of X values. In the latter case, the table shows the observed and critical difference for each pair of X values; significant differences are marked by ^*^ ^Graph intervals^ shows a graphic version of the table listing the pairwise con- fidence intervals or tests. For intervals, the graph shows horizontal er- ror bars ordered (from top to bottom) to match the corresponding table. For tests, the graph shows a triangular plot with symbols whose sizes reflect the magnitude of a difference in Y means, compared to its associated criti- cal difference. Non-significant differences are invisible; the axes of the plot match the rows and columns of the corresponding table ^Show legend^ if checked displays an explanatory legend near the top of the graph of confidence intervals or significance tests ^Graph options^ accepts most options available with the @graph@ ^twoway^ style. Ex- cluded options are ^yreverse^, ^ylabel^ and, for confidence intervals, ^b2title^ and ^xscale^; ^saving^ is also excluded: use the File/Save Graph dialog. Warnings -------- The left-quote (^`^) and double-quote (^"^) characters must be avoided in the ^^ and ^^ edit boxes, as well as in the ^Graph options^ edit box. Their presence can cause a variety of problems. Examples -------- See the on-line help for @prcomp@. Author ------ John R. Gleason Syracuse University Syracuse NY, USA loesljrg@@ican.net Also see -------- STB: STB-47 sg101 Manual: [R] graph; [R] oneway; [R] ttest On-line: ^help^ for @graph@, @oneway@, @ttest@ (@qsturng@, if installed)